What predicts preschoolers' executive functions a year later?
Preschoolers' own characteristics (earlier executive function, being a girl, higher family status, better fine motor skills and visual perception) predicted their executive functions a year later, while parenting, siblings and physical activity did not.
Source
Predictors of Executive Functions in Preschoolers: Findings From the SPLASHY Study
Study at a glance
- Design
- Cohort — Prospective multi-site cohort (SPLASHY): predictors measured at baseline in child care centres and parent questionnaires, executive function tested again one year later; analysed with multiple regression on 50 imputed datasets plus a lasso penalised regression
- N
- N=555 · 555 children entered the analysed sample; 476 were tested at baseline in 2014, 382 of them returned one year later, and 79 more were recruited in 2015 and had baseline data only. Missing values were multiply imputed.
- Population
- Typically developing Swiss children aged 2 to 6 recruited from 84 child care centres in five cantons, stratified by urban/rural and high/low SES
- Outcome
- Executive-function composite one year later: mean of age-adjusted scores on a selective-attention card sort, the NEPSY statue (motor inhibition) task and a visuo-spatial working-memory task
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
The researchers followed Swiss preschoolers recruited through child care centres and measured a wide set of possible predictors at baseline: demographic and biological factors (sex, preterm birth, family socioeconomic status, body fat, motor skills, accelerometer-measured physical activity), psychological factors (parent-rated hyperactivity, peer problems, prosocial behaviour, emotionality, a visual-perception task, baseline executive function) and interpersonal factors (parenting stress, positive and inconsistent parenting, siblings, time outdoors, days in child care). One year later the children were retested on three executive-function tasks tapping inhibition and working memory. All predictors were entered together in a multiple regression and, to guard against overfitting, a lasso model that shrinks weak predictors to zero.
What they found
Together the predictors explained about 31% of the variance in later executive function. The strongest predictor was executive function at baseline (beta 0.30), followed by smaller effects of fine motor skills (0.17), visual perception (0.16), socioeconomic status (0.15) and female sex (0.14). Parenting style, parenting stress, siblings, days in child care, time outdoors, physical activity, preterm birth and child temperament did not predict executive function, and the lasso shrank the parenting coefficients to zero.
The limits
What it doesn't show
This is observational, so even the significant predictors are not shown to cause better executive function, and the authors could not test mechanisms such as language skills that may link family status to executive function. There were only two time points one year apart, and baseline executive function was only moderately stable, so the young children's scores were noisy. The sample was healthy and physically active with ordinary parenting scores, which may explain why parenting and activity showed no effect; parenting was self-reported, and the peer-problems scale had poor reliability. The null results for parenting do not rule out effects in higher-risk families.
Key terms
- Executive functions
- Top-down control processes such as inhibition, working memory and cognitive flexibility that let a person plan and regulate behaviour.
- Lasso regression
- A penalised regression that shrinks the coefficients of weak predictors toward or exactly to zero, reducing overfitting when many predictors are tested at once.
- Multiple imputation
- Filling in missing data many times with plausible values, running the analysis on each completed dataset and pooling the results.
- Socio-ecological model
- A framework that sorts influences on development into nested levels, from the individual child to family and wider social context.
- Fine motor skills
- Precise, coordinated movements of the hands and fingers, such as manipulating small objects.
Flashcards
0 of 10 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
In this cohort, which factor was the strongest predictor of executive function one year later?
Common questions
Does this mean parenting doesn't matter for executive function?
Not necessarily. In this healthy, low-risk sample parenting stress and style were close to population norms and self-reported, so there may have been too little variation to detect an effect; earlier studies with observed parenting have found small effects.
Why use both ordinary regression and lasso?
With many predictors, ordinary regression can overfit the sample. The lasso gives up a little bias to gain predictive accuracy; predictors that survive the lasso are more likely to replicate, which is why the authors treat it as a check on the main model.
Why does baseline executive function only moderately predict later executive function?
At ages 2 to 6 executive functions are still developing quickly and measurement is noisy, so children's rank order changes a lot over a year.
More on Executive function